Jerry Wei

Google DeepMind (Google) researcher and first author of the long-form factuality paper. It introduced the LongFact prompt set, SAFE (an LLM agent that checks each claim with Google Search) and the F1@K metric.

Known For

• Long-form Factuality in Large Language Models (arXiv:2403.18802, NeurIPS 2024). SAFE agreed with crowd annotators on 72% of facts, was right in 76% of 100 sampled disagreements, and cost more than 20× less.

Articles

• Long-form Factuality in Large Language Models

Related Concepts

• Atomic Claim Verification
• Hallucination Detection

Related Datasets

• LongFact

People

• Sewon Min: FActScore, which SAFE builds on